D-Bus 1.7.0 (UNRELEASED)
==

Build-time configuration changes:

• The --with-dbus-session-bus-default-address configure option is no longer
  supported. Use the new --with-dbus-session-bus-connect-address and
  --with-dbus-session-bus-listen-address options instead. On Windows, you
  usually want them to have the same argument; on Unix, the defaults are
  usually correct.

• Similarly, the DBUS_SESSION_BUS_DEFAULT_ADDRESS cmake variable is no longer
  supported; use the new DBUS_SESSION_BUS_LISTEN_ADDRESS and
  DBUS_SESSION_BUS_CONNECT_ADDRESS variables instead.

Enhancements:

• D-Bus Specification 0.21
  · actually say that /org/freedesktop/DBus is the object that
    implements o.fd.DBus (fd.o #51865, Colin Walters)
  · various reorganisation for better clarity (fd.o #38252, Simon McVittie)
  · stop claiming that all basic types work just like INT32 (strings don't!)

Fixes:

• In the activation helper, when compiled for tests, do not reset the system
  bus address, fixing the regression tests. (fd.o #52202, Simon)

• Fix building with Valgrind 3.8, at the cost of causing harmless warnings
  with Valgrind 3.6 on some compilers (fd.o #55932, Arun Raghavan)

• Don't leak temporary fds pointing to /dev/null (fd.o #56927, Michel HERMIER)

D-Bus 1.6.8 (2012-09-28)
==

The "Fix one thing, break another" release.

• Follow up to CVE-2012-3524: The additional hardening
  work to use __secure_getenv() as a followup to bug #52202
  broke certain configurations of gnome-keyring.  Given
  the difficulty of making this work without extensive
  changes to gnome-keyring, use of __secure_getenv() is
  deferred.

D-Bus 1.6.6 (2012-09-28)
==

The "Clear the environment in your setuid binaries, please" release.

Fixes:

• CVE-2012-3524: Don't access environment variables (fd.o #52202)
  Thanks to work and input from Colin Walters, Simon McVittie,
  Geoffrey Thomas, and others.

• Unix-specific:
  · Fix compilation on Solaris (fd.o #53286, Jonathan Perkin)
  · Work around interdependent headers on OpenBSD by including sys/types.h
    before each use of sys/socket.h (fd.o #54418, Brad Smith)

• Windows-specific:
  · The default session bus listening and connecting address is now
    "autolaunch:", which makes D-Bus on Windows interoperate with itself
    and GDBus "out of the box". Use the configure options and cmake variables
    described above if you require a different autolaunch scope.
    (fd.o #38201, Simon McVittie)

D-Bus 1.6.4 (2012-07-18)
==

Fixes:

• Detect that users are "at the console" correctly when configured with
  a non-default path such as --enable-console-auth-dir=/run/console
  (fd.o #51521, Dave Reisner)

• Remove an incorrect assertion from DBusTransport (fd.o #51657,
  Simon McVittie)

• Make --enable-developer default to "no" (regression in 1.6.2;
  fd.o #51657, Simon McVittie)

• Windows-specific:
  · Launch dbus-daemon correctly if its path contains a space
    (fd.o #49450, Wolfgang Baron)

D-Bus 1.6.2 (2012-06-27)
==

The "Ice Cabbage" release.

• Change how we create /var/lib/dbus so it works under Automake >= 1.11.4
  (fd.o #51406, Simon McVittie)

• Don't return from dbus_pending_call_set_notify with a lock held on OOM
  (fd.o #51032, Simon McVittie)

• Disconnect "developer mode" (assertions, verbose mode etc.) from
  Automake maintainer mode. D-Bus developers should now configure with
  --enable-developer. Automake maintainer mode is now on by default;
  distributions can disable it with --disable-maintainer-mode.
  (fd.o #34671, Simon McVittie)

• Automatically define DBUS_STATIC_BUILD in static-only Autotools builds,
  fixing linking when targeting Windows (fd.o #33973; william, Simon McVittie)

• Unix-specific:
  · Check for libpthread under CMake on Unix (fd.o #47237, Simon McVittie)

D-Bus 1.6.0 (2012-06-05)
==

The “soul of this machine has improved” release.

This version starts a new stable branch of D-Bus: only bug fixes will
be accepted into 1.6.x. Other changes will now go to the 1.7.x branch.

Summary of changes since 1.4.x:

• New requirements
  · PTHREAD_MUTEX_RECURSIVE on Unix
  · compiler support for 64-bit integers (int64_t or equivalent)

• D-Bus Specification v0.19

• New dbus-daemon features
  · <allow own_prefix="com.example.Service"/> rules allow the service to
    own names like com.example.Service.Instance3
  · optional systemd integration when checking at_console policies
  · --nopidfile option, mainly for use by systemd
  · path_namespace and arg0namespace may appear in match rules
  · eavesdropping is disabled unless the match rule contains eavesdrop=true

• New public API
  · functions to validate various string types (dbus_validate_path() etc.)
  · dbus_type_is_valid()
  · DBusBasicValue, a union of every basic type

• Bug fixes
  · removed an unsafe reimplementation of recursive mutexes
  · dbus-daemon no longer busy-loops if it has far too many file descriptors
  · dbus-daemon.exe --print-address works on Windows
  · all the other bug fixes from 1.4.20

• Other major implementation changes
  · on Linux, dbus-daemon uses epoll if supported, for better scalability
  · dbus_threads_init() ignores its argument and behaves like
    dbus_threads_init_default() instead
  · removed the per-connection link cache, improving dbus-daemon performance

• Developer features
  · optional Valgrind instrumentation (--with-valgrind)
  · optional Stats interface on the dbus-daemon (--enable-stats)
  · optionally abort whenever malloc() fails (--enable-embedded-tests
    and export DBUS_MALLOC_CANNOT_FAIL=1)

Changes since 1.5.12:

• Be more careful about monotonic time vs. real time, fixing DBUS_COOKIE_SHA1
  spec-compliance (fd.o #48580, David Zeuthen)

• Don't use install(1) within the source/build trees, fixing the build as
  non-root when using OpenBSD install(1) (fd.o #48217, Antoine Jacoutot)

• Add missing commas in some tcp and nonce-tcp addresses, and remove
  an unused duplicate copy of the nonce-tcp transport in Windows builds
  (fd.o #45896, Simon McVittie)

D-Bus 1.5.12 (2012-03-27)
==

The “Big Book of Science” release.

• Add public API to validate various string types:
  dbus_validate_path(), dbus_validate_interface(), dbus_validate_member(),
  dbus_validate_error_name(), dbus_validate_bus_name(), dbus_validate_utf8()
  (fd.o #39549, Simon McVittie)

• Turn DBusBasicValue into public API so bindings don't need to invent their
  own "union of everything" type (fd.o #11191, Simon McVittie)

• Enumerate data files included in the build rather than using find(1)
  (fd.o #33840, Simon McVittie)

• Add support for policy rules like <allow own_prefix="com.example.Service"/>
  in dbus-daemon (fd.o #46273, Alban Crequy)

• Windows-specific:
  · make dbus-daemon.exe --print-address (and --print-pid) work again
    on Win32, but not on WinCE (fd.o #46049, Simon McVittie)
  · fix duplicate case value when compiling against mingw-w64
    (fd.o #47321, Andoni Morales Alastruey)

D-Bus 1.5.10 (2012-02-21)
==

The "fire in Delerium" release.

On Unix platforms, PTHREAD_MUTEX_RECURSIVE (as specified in POSIX 2008 Base
and SUSv2) is now required.

• D-Bus Specification 0.19:
  · Formally define unique connection names and well-known bus names,
    and document best practices for interface, bus, member and error names,
    and object paths (fd.o #37095, Simon McVittie)
  · Document the search path for session and system services on Unix, and
    where they should be installed by build systems (fd.o #21620, fd.o #35306;
    Simon McVittie)
  · Document the systemd transport (fd.o #35232, Lennart Poettering)

• Make dbus_threads_init() use the same built-in threading implementation
  as dbus_threads_init_default(); the user-specified primitives that it
  takes as a parameter are now ignored (fd.o #43744, Simon McVittie)

• Allow all configured auth mechanisms, not just one (fd.o #45106,
  Pavel Strashkin)

• Improve cmake build system (Ralf Habacker):
  · simplify XML parser dependencies (fd.o #41027)
  · generate build timestamp (fd.o #41029)
  · only create batch files on Windows
  · fix option and cache syntax
  · add help-options target
  · share dbus-arch-deps.h.in with autotools rather than having our
    own version (fd.o #41033)

• Build tests successfully with older GLib, as found in e.g. Debian 6
  (fd.o #41219, Simon McVittie)

• Avoid use of deprecated GThread API (fd.o #44413, Martin Pitt)

• Build documentation correctly if man2html doesn't support filenames on
  its command-line (fd.o #43875, Jack Nagel)

• Improve test coverage. To get even more coverage, run the tests with
  DBUS_TEST_SLOW=1 (fd.o #38285, #42811; Simon McVittie)

• Reduce the size of the shared library by moving functionality only used
  by dbus-daemon, tests etc. into their internal library and deleting
  unused code (fd.o #34976, #39759; Simon McVittie)

• Add dbus-daemon --nopidfile option, overriding the configuration, for
  setups where the default configuration must include <pidfile/> to avoid
  breaking traditional init, but the pid file is in fact unnecessary; use
  it under systemd to improve startup time a bit (fd.o #45520,
  Lennart Poettering)

• Optionally (if configured --with-valgrind) add instrumentation to debug
  libdbus and associated tools more meaningfully under Valgrind
  (fd.o #37286, Simon McVittie)

• Improve the dbus-send(1) man page (fd.o #14005, Simon McVittie)

• Make dbus-protocol.h compatible with C++11 (fd.o #46147, Marc Mutz)

• If tests are enabled and DBUS_MALLOC_CANNOT_FAIL is set in the environment,
  abort on failure to malloc() (like GLib does), to turn runaway memory leaks
  into a debuggable core-dump if a resource limit is applied (fd.o #41048,
  Simon McVittie)

• Don't crash if realloc() returns NULL in a debug build (fd.o #41048,
  Simon McVittie)

• Unix-specific:
  · Replace our broken reimplementation of recursive mutexes, which has
    been broken since 2006, with an ordinary pthreads recursive mutex
    (fd.o #43744; Sigmund Augdal, Simon McVittie)
  · Use epoll(7) for a more efficient main loop in Linux; equivalent patches
    welcomed for other OSs' equivalents like kqueue, /dev/poll, or Solaris
    event ports (fd.o #33337; Simon McVittie, Ralf Habacker)
  · When running under systemd, use it instead of ConsoleKit to check
    whether to apply at_console policies (fd.o #39609, Lennart Poettering)
  · Avoid a highly unlikely fd leak (fd.o #29881, Simon McVittie)
  · Don't close invalid fd -1 if getaddrinfo fails (fd.o #37258, eXeC001er)
  · Don't touch ~/.dbus and ~/.dbus-keyrings when running 'make installcheck'
    (fd.o #41218, Simon McVittie)
  · Stop pretending we respect XDG_DATA_DIRS for system services: the launch
    helper doesn't obey environment variables to avoid privilege escalation
    attacks, so mak